International Interactions, 34:402–422, 2008 Copyright © Taylor & Francis Group, LLC ISSN: 0305-0629 DOI: 10.1080/03050620802561462

The Politics of Foreign Direct Investment: An Interactive Framework

1547-7444 Interactions, 0305-0629 GINI International Interactions Vol. 34, No. 4, Nov 2008: pp. 0–0

Politics T. G. Coan of Foreign and T. Direct Kugler Investment

TRAVIS G. COAN Claremont Graduate University, Clarement, California, USA

TADEUSZ KUGLER Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

La Sierra University, Riverside, California, USA

Recent scholarship on the political determinants of foreign direct investment (FDI) point to the importance of a government’s level of political capacity. Governments with high levels of capacity, it is argued, have the political power and economic resources to carry out preferred policy objectives. This line of reasoning, however, fails to provide insight into what these policy objectives are likely to include. The present study attempts to overcome this deficiency in the literature by modeling explicitly the interactive relationship between open market policy environments and relative political capacity (RPC). We argue that governments with open market policy frameworks and high levels of political capacity send clear signals of a political environment conducive to sustained profitability. Our empirical results confirm the expected interactive relationship, providing important insight into the expected effects of open market policies and political capacity.

KEYWORDS foreign direct investment, political capacity, open market policy, interactive statistical analysis

Although the majority of private investment continues to come from domestic sources, time series data shows an enormous increase in FDI across both the developed and developing world (World Bank, 2005). Moreover, a number of studies show that foreign investment provides a method to incorporate new knowledge from abroad and facilitates technology spillovers (Mansfield and

Address correspondence to Travis G. Coan, 2993 N. Garey Ave. Pomona, California 91767, USA. E-mail: [email protected] 402

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Politics of Foreign Direct Investment

403

Romeo, 1980), suggesting that the impact of FDI on host economies may be greater than the sum of its parts. Given its implications for economic growth in developing economies (Borensztien, De Gregorio, and Lee, 1998), recent attempts in the literature to understand how countries attract FDI are of particular importance. With these potential economic benefits in mind, this study provides a thorough empirical examination of the link between political structures and the ability of nations to attract foreign investment. It is now widely accepted that political institutions—such as regulatory structures, macroeconomic policies, political rights, economic freedom, and private property rights—are critical for a nation’s ability to facilitate a positive investment climate (World Bank, 2004, 2005, 2008; Kaufmann, Kraay, and Mastruzzi, 2003; De Soto, 1989, 2000). Specifically, policy environments that allow for (1) the free movement of capital, (2) a stable investment environment, (3) the ability to start a business with relative ease, and (4) a set of institutionally protected rights that protect capital repatriation and prevent expropriation, provide an attractive climate for international investors. This study argues that a beneficial policy environment is not enough— governments must have the capacity or political power to actualize these environments. Governments that lack the capacity to mobilize diverse groups, the resources to provide critical economic infrastructure, and the bureaucratic capacity to follow through on policy promises are unlikely to signal the capability to provide and sustain open market policies. Thus, introducing the concept of political capacity allows us to estimate a more fully specified model of the political determinants of FDI. Consistent with recent work in political economy, we contend that political capacity allows elites to more effectively carry out their policy objectives. To the extent that capacity implies increases in private investment, “one important assumption of this argument is that the strong government in question should be growth oriented” (Feng, 2003, p. 167, emphasis in original). With respect to FDI, this implies that governments not only express a commitment to economic growth, but to the open market policies that facilitate foreign capital accumulation. The central claim of this study is that governments with open market policy frameworks and high levels of political capacity send clear signals to international investors of politically favorable investment environments. We postulate that governments must have both the means (political capacity) and the motives (a commitment to open market policies) for an economy to attract direct investment. To test this proposition, we develop an empirical model that transforms the common “growth oriented” assumption into an empirically testable proposition, allowing us to examine its consequences for predicting empirical realities. Using panel data for a sample of nations during the period 1995–2004, we test the conditional interactive relationship of policy environments and political capacity. Our empirical results support this insight. The interaction between open market policies and political capacity is an important

404

T. G. Coan and T. Kugler

determinant of FDI. This interactive relationship has a number of important policy implications for governments that wish to attract foreign investors.

LITERATURE REVIEW The empirical literature on the determinants of FDI has generally progressed along two different (though not entirely separate) dimensions—one economic and the other political. The following section provides an overview of the economic and political science perspectives used to construct our own theoretical framework.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

ECONOMICS Markets, Natural Resources, and Growth While the economic empirical literature on the determinants of FDI is too large to review in its entirety, we find that the most commonly researched measures include market size and the demographics of the population, the availability of natural resources, and the possibilities of economic growth. Moran et al. (2005) summarizes a number of core consistent findings that emerge from this scholarship. Much of the economic work is focused on empirically validating and extending John Dunning’s (1981) influential ownership, location, and internalization (OLI) model. Consistent with Dunning’s framework, a number of studies have isolated location-related factors as crucial for explaining cross-country variation in FDI (Schneider and Frey, 1985; Billington, 1999). Measures of market size—measured as aggregate or per-capita GDP— and population are among the most robust economic determinants of FDI flows (Chakrabarti, 2001; Wei, 2000; Feng, 2003).1 Theoretically, measures of market size and population provide a rough indicator of the profitability associated with supplying goods to a foreign market. Specifically, total population provides an upper bound on the number of consumers available in a host economy (Banga, 2003). Economic Growth is another indicator consistently included in empirical studies of FDI. The future potential of a nation’s economy has direct effects not simply upon the level of capital available at any given point in time, but also on changes in the economic incentive structures of the population, on the type of FDI available to a particular nation, and on overall investment opportunities (Chen and Mohsin, 1997; Billington, 1999). More specifically, changes in growth can both attract new FDI and alter the dynamics of older FDI, as the incentive structures of the population change in relation to the changing levels of income. Development leads not simply to increased customer bases, but to more important shifts within industries in host economies, which may in time spur future FDI (Wheeler and Mody, 1992).

Politics of Foreign Direct Investment

405

In addition to measures of economic growth, recent scholarship examines the link between foreign capital flows and natural resources. Theoretically, FDI will flow to areas with high concentrations of natural resources, as these areas allow for profitable private extractions. Indeed, research shows that even in periods of minimal international capital movement, direct investment to secure natural resources remains relatively constant and substantively important (Bravo-Ortega & de Gregorio, 2005). Moreover, given the high rates of return on natural resources extraction, investment tends to flow to areas that are politically risky and economically depressed (Ibid.). Thus, a nation’s level of natural resources provides an important source of foreign investment that is not captured by ‘traditional’ economic and political indicators.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

POLITICS While economic models provide a necessary baseline to study FDI, political factors are the primary focus of this study. Recent work on the politics of FDI pays particular attention to market policies that promote economic freedom and the level of political capacity. It is these variables that form the basis of our political study of FDI.2

Economic Freedom There is a growing literature on the relationship between economic freedom and private investment. This literature reports a number of consistent findings. First, policies that strengthen property rights, stabilize the court system, and improve the effectiveness of the public bureaucracy have been shown to have clear consequences for a firm’s ability—both domestic and foreign—to manage its investments (Hanson and Feestra, 2005; Acemoglu, Johnson, and Robinson 2005). Second, the cost of doing business is an important consideration for foreign investors and these costs are not uniform throughout the world (World Bank, 2004, 2008). Third, stringent regulatory environments, heavy tax burdens, and price controls all have detrimental effects on investment in general and FDI in particular (De Soto, 1989, 2000; Edwards, 1990). Lastly, a well-developed and liberal financial system is critical to the establishment of foreign capital. Research has shown that the inability to move money for investment purposes has detrimental effects for private investment and domestic capital accumulation (Foley, Mahir, and Hines, 2004).

Political Capacity Another political variable that has been explored in recent empirical studies on private investment—and the central variable in this study—is the

406

T. G. Coan and T. Kugler

concept of political capacity. In a study of the political and economic determinants of aggregate direct investment, Adji, Ahn, Holsey, and Willett (1997) report that relative measures of political reach and political extraction have significant and substantive positive effects on FDI in a series of crosscountry regressions. Feng (2003), in his extensive empirical study on political institutions, again finds a strong relationship between political capacity and direct investment. He reports that while political capacity has a modest impact on domestic private investment, it has statistically significant and substantively meaningful effects on attracting aggregate FDI stock.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

AN INTERACTIVE PERSPECTIVE ON THE POLITICS OF FDI How does politics impact the location decisions of international investors? To answer this question, we examine the specific policy environments that promote economic profitability and analyze a government’s capacity to actualize these environments. Before specifying the details of this relationship, it is important to define precisely what is implied by open market policies and political capacity. The foundation of open market policy environments is a nation’s institutional commitment to economic freedom. Following Beach and Kane (2007), we define economic freedom as the “right of property ownership; fully realized freedoms of movement for labor, capital, and goods, and an absolute absence of coercion or constraint of economic liberty beyond the extent necessary to maintain and protect liberty itself” (p. 40). In the context of FDI attraction, open market policies define the political infrastructure that allows international investors to freely move capital into a host economy, while providing the property rights to create foreign-owned enterprises, and to secure a reasonable return on investment. We postulate that a commitment to open market policies is not enough to attract international investment. Governments must have the political capacity or power to actualize a given set of policy commitments. We define political capacity as “the ability of political systems to carry out the tasks chosen by the nation’s government in the face of domestic and international groups with competing priorities” (Arbetman & Kugler, 1997, p. 12). Theoretically, political capacity refers to the ability of governments to reach into domestic societies and extract the resources necessary to facilitate a chosen policy. In the context of FDI, high levels of political capacity imply that governments have the capability to provide the physical infrastructure, political institutions, and mobilize the human resources necessary to ensure a profitable investment environment. To refresh: the central claim of this article is that governments with open market policy frameworks and high levels of political capacity send clear signals to international investors of politically favorable investment

Politics of Foreign Direct Investment

407

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

environments. Politically capable governments have the power to effectively implement and sustain economic policies which facilitate profitable investments, providing the basic political infrastructure to “do business.” However, policy and capacity must go together. Political capacity alone, when governments provide no indication as to the likely direction of their actions, sends a neutral or negative signal to international investors. Support for open market policy environments is an essential piece of the puzzle, providing valuable information on direction of political action. Specifically, support for open market policies allow investors to anticipate the actions governments can take, and predict the actions they will take. This argument suggests an interactive relationship between political capacity and economic policy and the following hypothesis: Governments attract FDI when they have the political capacity to realize their commitments to open market policies.

To our knowledge, no other study has attempted to estimate explicitly the interactive effects of political capacity and open market policy environments. The policy implications of understanding the proposed conditional relationship are potentially quite meaningful. If this interactive expectation holds, increasing FDI to politically capable governments without acknowledging the conditional effects of open market policies could lead foreign investors, unknowingly, into very risky environments.

DATA Scope of Study To test the relation between politics, economics, and FDI, we examine panel data on 103 developed and developing nations over the period 1995– 2004. Problems associated with missing data are endemic to multivariate empirical studies of FDI and the common method of list-wise deletion often leads to unrepresentative samples of the international system. Recent advances in the statistical literature suggest a number of methods to deal with the “missing data problem” and advances in computational power have provided researchers with the ability to transform earlier proposed “solutions” into useable computer applications (Allison, 2002). In order to avoid the pitfalls of list-wise deletion and maintain as much information as possible to examine the correlates of FDI, we employ the multiple imputation (MI) method described in King and Honaker (2005), which builds on earlier work from King et al. (2001).3 King and Honaker’s (2005) method uses a modified bootstrapping algorithm to accomplish the common steps associated with MI methods: 1) drawing estimates of means and variances from their posterior densities, 2) using these estimates to simulate the

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

408

T. G. Coan and T. Kugler

parameters of interest, and 3) using the simulated parameters to deterministically compute the imputed values (see Allison, 2002 for an accessible introduction to MI methods and King and Honaker, 2005 for a detailed explanation of the algorithm used in this paper). More important for the present study, King and Honaker’s algorithm provides a number of modifications that make it suitable for time series cross-sectional data. After implementing the algorithm, we analyze a total of 1030 observations, which provide a representative sample of the international system. While considerable time went into ensuring that our sample is representative, potential limitations exist: there is literature that suggests a differential impact of political indicators across the developed and developing world. Within the RPC literature, studies such as Arbetman and Kugler (1997) and Kugler and Swaminathan (2006) show that capacity displays a diminishing rate of return upon indicators such as income, births, deaths, and FDI when nations make the transition from developing to developed.4 Highly organized and capable countries have few untapped methods or structures to increase already high levels of capacity. Because of the limits associated with developing further capacity, we see the greatest effect of capacity at the lowest level of development. To explore the possibility of differential impacts across developed and developing societies, we not only analyze the entire sample, but also conduct separate analysis on a sample that excludes OECD countries. It is important to be specific regarding the policy implications of this study, and this adjustment permits the formation of future recommendations for both the developed and developing worlds. Since the consequences of combining policy and capacity are not evenly distributed across the world, our analysis must take into account these potential differences.

Dependent Variable: FDI As shown in previous scholarship (Borensztien, De Gregorio, and Lee, 1998), FDI provides an important source of fixed-capital formation for both developing and developed economies. Investment in the form of fixed capital is among the most robust predictors of economic growth, while savings and investment are key components in classical and endogenous growth theories (Schmidt, 2003). Given the emphasis on fixed-capital formation, and consistent with a number of studies on the politics of FDI (c.f., Wei, 2000; Feng, 2003), we utilize FDI stock within a host country. Focusing on direct investment stocks rather than flows captures both the dynamic changes in internationally owned investment capital and highlights the cross sectional, aggregate differences among countries that have accumulated FDI overtime. Moreover, FDI stock effectively captures a country’s medium to long-term equilibrium level of international capital, and reflects differences across these values. When studying the politics of FDI, it is these equilibrium

Politics of Foreign Direct Investment

409

values that are of primary interest because political indicators are generally subject to slow, evolutionary changes and thus cannot be expected to track quarter-to-quarter fluctuations in FDI flows. Data on FDI stock comes from the United Nations Conference on Trade and Development’s (UNCTAD) FDI database and is measured in millions of U.S. dollars.

Independent Variables

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Our basic proposition is that capable governments with open-market policy environments will attract high levels of FDI. Falsification of this proposition requires two core pieces of information: objective measures of political capacity and open market policy environments. POLITICAL

CAPACITY

Political capacity is approximated by the measure of relative political capacity (RPC) developed in Arbetman and Kugler, 1997. (See Ch. 2 for a detailed overview of the indicator’s construction.) RPC has been used constructively to link political institutions to a wide range of economic indicators, including economic growth, inflation, and private investment. The measure is constructed using a series of multiple regression models to predict the expected levels of extraction, and compares the expected levels to actual levels of extraction. RPC is an objective measure of government extraction that captures the disposable resources a government has to advance stated policy goals. The measure is constructed such that higher values indicate greater levels of government capacity, with a mean of 1.02—i.e., the normal rate of extraction—and range from .029 to 2.29 in our sample. OPEN

MARKETS

While it is generally agreed that open market policies are important drivers of private investment, the challenge of measuring this concept is not trivial. Extant empirical literature on open market policy environments tends to rely on one of two sources of information to construct indicators of economic freedom: expert opinions or factual data based on financial records and legal structures. Although there are problems associated with using both types of information, as much as possible, the empirical work that follows relies on factual data. The advantage of using objective information to create an indicator of open market policies is the ability to enhance the validity of cross country comparisons, while attempting to minimize subjectivity in the data generation process. Specifically, to construct a measure of open market environments, this study utilizes six measures of economic freedom from the Heritage Foundation—investment freedom, business freedom,

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

410

T. G. Coan and T. Kugler

fiscal freedom, monetary freedom, trade freedom, and property rights (see Beach and Kane, 2007 for the methodology used to construct each indicator). Each Heritage measure, as much as possible, utilizes objective information from a number of sources and these measures have been used productively in a wide range of empirical analysis on the causes and effects of economic freedom (Mlambo, 2006, Steinberg & Saideman, 2008).5,6 These measures operationalize the ability to move investment capital (and products) into and out of host nations, while also incorporating the political rights necessary to protect private property and investment capital. This specification integrates the multidimensional and interactive nature of open market policies suggested in other work (La Porta et al., 1997, 1998; King and Levine, 1993). The multidimensional nature of policy environments presents the challenge of finding a suitable method to combine the underlying components. We use factor analytic techniques to combine the six Heritage measures into a single Policy scalar. Specifically, we conducted a principal components factor analysis for each of the freedom variables.7 As expected, only one factor with an eigenvalue over 1 emerged and all six variables loaded highly on the single dimension in the same direction—the rotated loadings ranged from a high of .84 (property rights) to a low of .58 (monetary freedom), with most variables over .80. After extracting the factor, we used regression scoring to produce a single indicator and, in order to ease interpretation, rescaled the variable to range between 0 and 1. The Policy measure is only marginally correlated with RPC (r = .32), and thus one may be confident that Policy is measuring a dimension of political choices that is independent from government capacity. CONTROL

VARIABLES

In order to avoid the well-known problem of omitted variable bias, a number of economic controls variables were used.8 While the economics literature suggests a wide range of potential correlates, our review of the literature indicates that only market size shows a robust relationship with FDI capital accumulation (Chakrabarti, 2001). To control for the potential effects of market size, we include two measures, the natural logarithm of population and a nation’s level of economic development which is measured as per-capita GDP at purchasing power parity (PPP). In addition to including these base variables, we include controls for a nation’s level of economic growth (measured as the percent change in per-capita GDP), as well as their level of “natural resources” (proxied as agriculture exports plus mining exports). Each variable is discussed regularly in the extant economic literature on FDI and thus warrant inclusion in the model. Data on population, per-capita GDP, and the data used to construct the natural resources measure comes from the World Development Indicators.

Politics of Foreign Direct Investment

411

MODEL SPECIFICATION AND RESULTS To test the argument presented in Hypothesis 1, we estimate an econometric model of the following form:

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

FDI it = b0 + b1 RPC it + b2 Policyit + b3 RPC it * Policyit + aC it + eit

(1)

Where FDI represents the stock of foreign direct investment inflows; RPC represents a measure of relative political capacity; Policy represents a proxy for open market policy environments; RPC * Policy represents the interaction between government capacity and open market policy; C represents a vector of control variables; and ∈ represents the stochastic error term. The subscripts in (1) denote country i in time t and thus incorporate the model’s panel structure. The interaction between RPC and Policy, as well as the additive inclusion of RPC and Policy, allows us to examine directly the conditional relationship specified above. This specification permits us to analyze the differential impacts of RPC at various levels of the conditioning variable Policy and vice versa.

Econometric Considerations The use of panel data opens up a number of potential statistical problems; the most common include contemporaneous correlation of errors, cross panel heteroskedasticity, and autocorrelation. Preliminary statistical analysis suggested that these factors may be affecting the validity of our empirical results. To adjust for contemporaneous correlation and cross panel heteroskedasticity, we followed Beck and Katz (1995) and estimated “panel corrected standard errors.” Preliminary analysis also suggested the presence of a first-order autoregressive process in FDI stock inflows. The presence of an AR(1) process was anticipated given the way measures of FDI stock are constructed. Again, we follow Beck and Katz (1995) and adjust for the presence of first-order autocorrelation using a Prais-Winston transformation of the error structure. Lastly, to ensure that our results are not biased by problems related to endogenity, we used lagged variables of per-capita GDP and economic growth in the results that follow. In addition to the econometric considerations related to using crosssectional time series data, it is necessary to adjust the coefficient estimates and standard errors to reflect the uncertainty associated with using imputed data. To accomplish this objective, we followed the basic approach advocated in Rubin (1987) (see also Allison, 2002). First, we imputed 5 separate datasets to using King and Honaker’s (2005) procedure. Second, we used each of the five datasets to estimate (1)—taking the econometric considerations above into consideration—and averaged across the datasets to derive

412

T. G. Coan and T. Kugler

the coefficient estimates for each variable presented below. Lastly, we calculate the adjusted panel corrected standard error for each coefficient estimate using the following formula (see Allison, 2002):

SE ( b ) =

1 M



1 

1



∑ sk 2 +  1 + M   M − 1 ∑ ( bk − b )2

(2)

k

Where b is the estimated coefficient, M represents the number of replications, sk represents the estimated standard error of replication k, and βk represents the kth estimate of the parameter β.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Empirical Results Table 1 presents the empirical results for the interactive statistical model specified in (1), including the adjusted nonstandardized regression coefficients and panel corrected standard errors adjusted to reflect (2). To examine our core propositions, we present four empirical models. Model 1 examines the independent effects of the economic control variables for the entire sample, while Model 2 examines the effects of the key political variables for the entire sample, including the interactive effects of RPC and Policy. Model 3 and Model 4 repeat the analyses in the first two models, while restricting the sample to include only developing countries. TABLE 1 Empirical Results

Variable RPC Policy RPC*Policy Log(Population) Per-capita GDPt-1 Growtht-1 Natural Resources Constant R2 F-Prob

Model 1 (Entire Sample)

Model 2 (Entire Sample)

Model 3 (Developing Nations)

Model 4 (Developing Nations)

– – – – – – 12631*** 4932.43 2.89*** 1.04 −17266 10577.79 1.10E-07*** 2.79E-08 −2.20E + 05*** 82177.03 0.37 0.00

−28279** 13382.24 −34771 25420.73 55266* 28629.38 12343*** 4527.202 2.57*** 0.90 −20874* 11816.34 1.20E-07*** 2.67E-08 −2.00E + 05*** 76379.37 0.40 0.00

– – – – – – 3255.9*** 1140.57 1.4651*** 0.44 −9596** 4504.22 6.50E-08*** 1.40E-08 −55710*** 18971.8 0.42 0.00

−19647* 11382.41 −31796* 17920.43 44863** 22529.23 3442.8*** 987.2504 1.44*** 0.36 −14803*** 5669.844 6.80E-08*** 1.23E-08 −46255** 18185.97 0.51 0.00

*p < .10; **p < .05; ***p < .01.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Politics of Foreign Direct Investment

413

Turning first to the traditional economic control variables, we find consistent results across each sample. Table 1 indicates that a nation’s level of economic development, natural resources, and the natural logarithm of population are significant at the 1% error level, with each variable displaying the correct theoretical sign. These results are generally consistent with the extant literature on the correlates of FDI stock inflows (c.f. Feng, 2003). Economic growth, on the other hand, fluctuates from insignificant to significant depending on the specification analyzed and consistently displays a sign opposite of theoretical expectations. The effect of economic growth on FDI has long been the subject of controversy in empirical studies of FDI. While a number of studies report positive findings, the results appear sensitive to the period under study (c.f. Tsai, 1994) and finding a negative relationship is not without precedent in the literature (c.f. Nigh, 1988). While more research is certainly needed to flesh out the effects of growth, overall the economic variables provide a relatively good explanation of variation in FDI stock inflows. More important for the theoretical question posed in this study, however, is the conditional relationship between government capacity and open market policies. The results are consistent with theoretical expectations. As shown in Table 1, the RPC*Policy interaction term and the additive term for RPC display the correct theoretical signs (positive and negative, respectively), and the results hold with and without the inclusion of OECD countries. Moreover, both variables are significant at traditional levels across models. These findings collectively support our original proposition—that is, capable governments with open-market policies are able to use political power to further attract direct investment. Governments that have the political strength and monetary resources to follow through on open-market policy promises, provide an important signal to international investors regarding their willingness to use political tools to attract and preserve investments. This signal is ineffective when policy environments are not in line with basic economic freedoms and rights (as indicated by the negative sign for the additive RPC term). Regarding the additive impact of Policy, we also find consistent evidence across the various specifications. For both the international system overall and the developing world in particular, we find a negative relationship between the additive impact of Policy and FDI inflows—though only for developing countries is the relationship significant at traditional levels. This finding further supports the proposition that a beneficial policy environment is not enough—governments must have the capacity and political power to make effective use of a good policy environment. Open-market policies in isolation provide little information from which to judge the effectiveness of current policy regimes and to forecast the likely direction of future regulations. Moreover, attempts to implement policy environments committed to economic freedom without the requisite political power may even reduce the level of FDI inflows, at least in developing countries. In short, weak governments

414

T. G. Coan and T. Kugler

with open market policies fail to send credible signals of a political commitment to securing a profitable investment environment. Nevertheless, it is important not to misinterpret the results presented in Table 1. The determination of statistical significance and the correct marginal effects for interactive models requires calculating the standard errors and coefficient values of the variable of interest at multiple levels of the conditioning variable (see Brambor, Clark, and Golder, 2006 for an overview). To ensure proper estimation, the results in Table 1 are supplemented with Figure 1. Figure 1 presents the significance level and coefficient estimate of RPC (Policy) at the minimum, mean, and maximum values of Policy (RPC) for both the 30000 23,108 **

Coefficient for RPC

20000 10000 3,941

6,508 *

0 Min

Mean

Max

–10000 –20000

–15,340 * –22,973 **

–30000 Policy Entire Sample

Non OECD

100000 84,603 **

80000

Coefficient for Policy

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

24,390

65,108 **

60000 40000 22,153 ** 14,413 **

20000 0 Min

Mean

Max

–20000 –40000

–30,450 –33,113 *

RPC Entire Sample

Non OECD

*p < .10; **p < .05; ***p < .01

FIGURE 1 Conditional Effects of RPC and Policy.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Politics of Foreign Direct Investment

415

entire sample and developing nations. The estimated coefficients and standard errors are thus based on Model 2 and Model 4, respectively. As shown in the top half of the figure, the estimated coefficients for RPC display a positive relationship based on the conditional effects of Policy—that is, RPC has a strong positive effect on direct investment at high levels of Policy and a strong negative effect at low levels. At the minimum level of Policy, the results are statistically significant for both samples. Specifically, these results show that when Policy is at its minimum level, a one-unit change in RPC leads to approximately a $23 billion dollar decrease in the level of FDI stock for a host economy in the sample including the OECD countries and a $15 billion dollar decrease in the developing world. Note that the relatively large effects of the coefficient estimates are due to the fact that RPC only ranges from .029 to 2.29. A more meaningful change in RPC will help to clarify this point. Figure 1 shows that if the average nation in our sample increases its level of RPC by 10%, say from a mean value of 1.02 to 1.12, the model predicts a roughly $2.3 billion dollar decrease in FDI stock for the entire sample and $1.5 billion dollar decrease for developing nations. These data strongly support the claim that political strength provides a negative signal to international investors in the absence of a commitment to open market policy frameworks. Consistent with our theoretical perspective, the opposite is true when governments effectively signal the capability to actualize a given set of open market policies. When Policy is at its maximum level, the results are statistically significant for developing nations and just outside of traditional significance levels for the entire sample (p = .11). For developing nations, a one-unit increase in RPC leads to a roughly $23 billion dollar increase in the level of FDI stock. Again, a more meaningful 10% increase in the value of RPC leads to a roughly $2.3 billion dollar reduction in the expected level of FDI. Moreover, for developing nations, a one-unit change in RPC, when Policy is held constant at its mean level, leads to a 6.5 billion dollar increase in FDI. This finding is also significant at the 10% error level. The combination of these findings provides support for the proposition that developing nations increase their ability to attract FDI when they have the political strength to sustain and effectively implement open market policy environments. The bottom half of Figure 1 provides the significance levels and coefficient estimates for Policy, taking into account the conditional effects of RPC. The interpretations of the coefficient values for Policy are the same as for RPC in the top half of the figure. Once more, the results display the same general pattern shown for RPC above: there is a clear positive relationship between the conditioning variable and Policy. At the minimum level of RPC, the effects for the Policy variable are insignificant for the entire sample and significant for developing countries. The results are also substantively meaningful, with a one-unit change leading to a roughly $33 dollar decrease in FDI stock. Again, these enormous substantive effects are due to the fact that

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

416

T. G. Coan and T. Kugler

Policy ranges from 0 to 1 and thus a one-unit change is equivalent to moving the variable from its minimum to maximum level—still, the substantive results are quite large. These findings suggest that open market policies provide little leverage in attracting FDI when government capacity is low and may actually reduce the level of FDI in developing nations. However, the results change substantially as government capacity increases. When government capacity is at its mean level, Policy has a positive effect for both samples and the results are significant at the 5% error level. Specifically, a one-unit change in Policy leads to a $14 billion dollar increase in FDI for the entire sample and $22 billion dollar increase for the developing world. The results are even stronger when RPC is at its maximum level, while the results maintain significance at the 5% error level. A one-unit change in Policy leads to a $65 billion dollar increase in FDI for the entire sample and $85 billion dollar increase for developing countries. Even with a more reasonable 10% change in Policy, the massive substantive effects remain and the findings are particularly strong for developing nations. Once again, the results provide additional evidence in support of our argument: government policies attract FDI when they have the political capacity to realize their commitment to a profitable investment environment. The evidence provided in Table 1 and Figure 1 is clear. First, strong governments with favorable economic policy environments are more apt to attract foreign capital. These governments signal a commitment to economic growth and provide valuable information to investors regarding the expected direction of economic policies. Second, the evidence provides strong support for a need to understand the conditional relationship between political capacity and open market policies. As shown in Figure 1, RPC and Policy could have a positive or negative effect on the accumulation of foreign capital and which is more likely depends on the level of the conditioning variable. Third, while the argument generalizes across samples, the estimated effects are generally stronger for developing countries. Each of these findings leads to important policy implications and it is to these implications that we turn in the next section.

DISCUSSION AND CONCLUSIONS We have provided a wealth of evidence to support our central proposition: governments attract FDI when they have the political capacity to realize their commitments to open market policies. These results generalize across nations in the international system, though the model appears to more effectively explain variation in FDI among developing nations. Moreover, the interactive perspective developed in this study allows one to examine explicitly the conditions under which political capacity and open market policies facilitate the accumulation of direct investment. These conditions are summarized in more detail as follows:

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Politics of Foreign Direct Investment

417

1. Capable governments with a strong commitment to open market policies send a clear signal to international investors regarding the nation’s sustained commitment to facilitating a profitable investment environment and the results hold across samples. Evidence of this claim is found in the positive sign on the interaction term in Table 1, as well as in the positive, statistically and substantively significant conditional effects presented in Figure 1. Collectively, these results suggest that capable governments with a commitment to basic economic freedoms are in the best position to attract a large share of the available foreign investment. 2. Incapable governments with open market policies fail to signal a political climate conducive to sustained profitability. Support for this claim is found in the insignificant additive effects for Policy in Table 1 for the entire sample, as well as the significant negative effects for developing countries. The results in Figure 1 further confirm the findings presented in Table 1, with incapable governments either displaying insignificant or negative effects depending on the sample analyzed. These results support the claim that economic freedom is not enough—governments must have the capacity to realize stated policy environments. 3. Capable governments without open market policies signal the ability to use their political power to promote policies inimical to foreign investment, potentially reducing economic growth. Evidence of this claim is provided by the negative additive effect for RPC in Table 1. Furthermore, as shown in Figure 1, when Policy is at its minimum level, the coefficient value for RPC is negative and statistically significant across each of our samples. 4. Incapable governments that lack a commitment to open market policies fail to provide investors with information on the expected direction of economic policy and thus reduce their ability to attract FDI. Evidence of this claim is provided in Figure 1 and is shown by the clear positive association present in the nonlinear relationship between RPC and Policy. These empirically validated conditions not only clarify the theoretical link between capacity, policy, and FDI, but also provide important safeguards against making erroneous policy proposals. Under the common “growth oriented assumption” (Feng, 2003), past empirical findings suggest that raising a nation’s level of political capacity will promote the attraction of FDI. A more nuanced view of political capacity, which takes into account the nonlinear effects of open market policies, suggests that raising capacity could actually reduce a nation’s level of FDI. Strong governments that lack a commitment to open market policies may use their power to further political objectives at the expense of economic objectives (Bueno de Mesquita et al., 2003). Thus, the conditional relationship between Policy and RPC provides the missing link between government capacity and the accumulation of foreign capital—that is, it allows interested parties to predict when government action is likely to spur foreign

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

418

T. G. Coan and T. Kugler

investment. Such information is useful to governments analyzing the investment potential of selected countries, particularly those in the developing world. The interactive relationship also provides insight into questions related to when and why measures to enhance basic economic freedoms fail to translate into private investment. These questions seem to be of particular importance given the recent inability of neoliberal market reforms to garner sustained acceptance across the developing world. Our model suggests that one explanation of the inability of neoliberal reforms to “take hold” may be found in taking a closer look at government capacity. Specifically, if governments lack the capability to provide necessary infrastructure and the ability to “mobilize human resources” in support of open market reforms, the model suggests that attempting to increase economic freedoms may actually decrease a nation’s ability to attract foreign investment. Thus, the interactive framework allows interested parties to identify the nations most likely to benefit from increasing marketoriented reforms and to predict when these reforms are likely to be ineffective. The macro-level findings in this study provide a useful starting point for studies on the micro-dynamics of open market policies, political power, and political signals of economically viable investment opportunities. In future research, we will explore this micro-level dynamic in provincial studies of India and China—two developing economies with arguably the largest investment potential. We propose to use our interactive framework to identify the provinces within nations with the political foundation to become successful hubs for foreign investment. This paper suggests the benefits of exploring explicitly the conditional hypotheses suggested by extant political-economy theories. The need to explore the nonlinearities inherent in political-economy interactions stretches, in our view, beyond the study of foreign investment. Similar arguments could offer insight into many topics in politics and economics, including economic growth, demography, and conflict. The policy implications will no doubt be fully taken advantage of by investors and may also provide guidance to policymakers charged with advancing national security.

NOTES 1. Chakrabarti found that in addition to variables related to market size, trade openness and wages rates provide robust predictors in cross country FDI regressions. The concept of trade openness is effectively incorporated into the model through the Policy indicator discussed below. Unfortunately, with respect to wages, the limited availability of data—particularly for developing nations—precluded us from examining its effects explicitly in the empirical model that follow. 2. Political instability is another indicator that has gained considerable attention in the political science scholarship. While some studies confirm the expected negative relationship (Schneider and Frey, 1985), others fail to find statistical significance (Asiedu et al., 2000). In general, the results seem to be sensitive to the indicator used to capture instability and the set of control variables used in the empirical analysis (Chakrabarti, 2001). In preliminary empirical work, political stability (measured using ICRG-data) failed to reach statistical significance at traditional levels and thus was included to simplify the model’s presentation. 3. The stand-alone Windows application Amelia was used to carry out the multiple imputation process (http://gking.harvard.edu/amelia/). The program was instantiated using the variables described

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Politics of Foreign Direct Investment

419

below, as well as their lagged values—i.e., FDI, RPC, Policy, RPC*Policy, Per-capita GDP, GDP Growth, Population, and Natural Resources. Log transformations of population, natural resources, and per-capita GDP were carried out to more closely approximate the MI model’s multivariate normal assumption. Moreover, as suggest in King and Honaker (2005), we ran a number of diagnostic checks on the imputed data, which confirmed a relatively good fit between the imputation model and observable data. In the end, the imputation process increased the number of useable observations in our dataset from 902 to 1030 or increased the sample size by roughly 13%. For access to the imputed data files and the source data used to carryout the imputation process, please contact the authors. 4. Kugler & Swaminathan (2006) show that RPC has the strongest influence on income, birth and death rates, when GDP per Capita is below $4,500 (pegged at 1992 U.S. dollars). 5. Each of these components is generated by combining data from a number of sources, including the World Bank, the International Monetary Fund, the U.S. Department of Commerce, Economist Intelligence unit, and the Office of the U.S. Trade Representative. For instance, to create the “investment freedom” variable used in this analysis, researchers scanned official reports from the above sources to examine questions such as whether there is a foreign investment code that defines the countries investment laws and procedures, whether there are government restrictions on foreign exchange, whether foreign firms are treated the same as domestic firms under the law. Researchers then define a rubric to place each nation on a 100-point scale, where 100 indicates the highest level of investment freedom (again, see Beach and Kane, 2007 for the methodology used to construct each indicator). 6. Alternative measurements of Economic Freedom such as those developed by Gwartney Lawson, and Leeson (2007) have also been used extensively in the investigation of the interconnection between economic freedoms and development. 7. The primary purpose of PCA is data reduction. Thus, the result of PCA is fewer variables, with less overlap and higher reliability between the variables. Varimax rotation was used to create more interpretable results, through maximization of the variance of the factor loadings. It is important to note that we experimented with alternative factor analytic methods—the results were similar in all cases, with a single component emerging with an eigenvalue over 1. Cronbach’s alpha indicates that the scale is reliable by traditional standards (alpha = .69). 8. In addition to the control variables described this section, we estimated models that controlled for a nations level of democracy (measured as Polity 2), the natural logarithm of GDP, and a measure of labor productivity (GDP/Labor Force). In all cases, the variables provided little explanatory power outside of the control variables included and thus to simplify the presentation (and reduce possible collinearity in the independent variables) we omitted these variables from the empirical models presented below.

REFERENCES Acemoglu, D., S. Johnson, and J. Robinson, (2005). “Institutions as the Fundamental Cause of Long-run Growth,” NBER wp 10481, Handbook of Economic Growth, North-Holland. Elsevier Adji, S.S., Y.S. Ahn, C.M. Holsey, and T.D. Willett (1997). Political Capacity and Capital Flows. In M., Arbetman, and J. Kugler, Political Capacity and Economic Behavior, Boulder, CO: Westview Press. Allison, P.D. (2002). Missing Data, Thousand Oaks, CA: Sage Publications, Inc. Arbetman, M. and J. Kugler (1997). Political Capacity and Economic Behavior. Boulder, CO: Westview Press. Asiedu, E. (2002). “On the Determinants of Foreign Direct Investment to Developing countries: Is Africa Different?” World Development, Vol. 30, No. 1, pp. 107–19. Banga, R, (2003). “Impact of Government Policies and Investment Agreements on FDI Inflows. Indian Council for Research on International Economic Relations, New Delhi Working Papers 116, Indian Council for Research on International Economic Relations, New Delhi, India.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

420

T. G. Coan and T. Kugler

Beach, W.W. and T. Kane (2007). Methodology: Measuring the 10 Economic Freedoms. Washington, DC: The Heritage Foundation. Beck, N. and J. Katz (1995). “What to Do (and Not to Do) with Time-Series CrossSectional Data in Comparative Politics.” American Political Science Review, Vol. 89, No. 3, pp. 634–47. Billington, N. (1999). “The Location of Foreign Direct Investment: An Empirical Analysis.” Applied Economics, Vol. 31, pp. 65–76. Borensztien, E., J. De Gregorio, and J-W. Lee (1998). “How Does Foreign Direct Investment Affect Growth?” Journal of International Economics, Vol. 45, pp. 115–135. Brambor, T., W.R. Clark, and M. Golder (2006). Understanding Interaction Models: Improving Empirical Analyses. Political Analysis Vol. 14, No. 1, pp. 63–82. Bravo-Ortega, C., and J. De Gregorio, (2005). The Relative Richness of the Poor? Natural Resources, Human Capital, and Economic Growth. Policy, Research working papers; No. WPS 3484. The World Bank. Bueno de Mesquita, Bruce, Alastair Smith, Randolph M. Siverson, and James D. Morrow (2003). The Logic of Political Survival, Cambridge, MA: MIT Press. Chakrabarti, A. (2001). “The Determinants of Foreign Direct Investment: Sensitivity Analyses of Cross-Country Regressions.” Kyklos, Vol. 54, No. 1, pp. 89–113. Chen, Z. and K. Mohsin (1997). “Patterns of Capital Flows to Emerging Markets: A Theoretical Perspective,” IMF Working Paper WP/97/13. International Monetary Fund, Washington D.C. De Soto, H. (1989). The Other Path: The Invisible Revolution in the Third World, New York: Harper & Row. De Soto, H. (2000). The Mystery of Capital: Why Capitalism Triumphs in the West and Fails Everywhere Else, New York: Basic Books. Doucouliagos, Chris and Mehmet Ali, Ulubasoglu (2006). “Economic Freedom and Economic Growth: Does Specification Make a Difference?” European Journal of Political Economy, Vol. 22, pp. 60–81. Dunning, J.H. (1981). International Production and the Multinational Enterprise, Boston: Allen and Unwin. Edwards, S. (1990). “Capital Flows, Foreign Direct Investment, and Debt Equity Swaps in Developing Countries,” Working Paper No. 3497, National Bureau of Economic Research, Cambridge, MA. Fana, J. P.H., R. Morckb, Lixin Colin Xuc, and B. Yeung (2007). “Does ‘Good Government’ Draw Foreign Capital? Explaining China’s Exceptional Foreign Direct Investment Inflow.” World Bank Policy Research Working Paper 4206. http:// econ.worldbank.org Feng, Y. (2003). Democracy, Governance, and Economic Performance: Theory and Evidence, Cambridge: The MIT Press. Foley, C.F. A.D. Mahir, and J.R. Hines Jr. (2004). “A Multinational Perspective on Capital Structure Choice and Internal Capital Markets.” Journal of Finance, Vol. 59, No. 6, pp. 2451–2488. Gwartney, J., R. Lawson, and R.G. Holcombe (1999). “Economic Freedom and the Environment for Economic Growth. Journal of Institutional and Theoretical Economics, Vol. 155, pp. 643–663.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

Politics of Foreign Direct Investment

421

Gwartney, J., R. Lawson, and P. T. Leeson (2007). Economic Freedom of the World: 2007 Annual Report. Vancouver, BC: The Fraser Institute. http://www. freetheworld.com Hanson, G., and R. Feestra (2005). Ownership and Control in Outsourcing to China: Estimating the Property Rights Theory of the Firm. Quarterly Journal of Economics, May 2005, Vol. 120, No. 2, pp. 729–762. Helpman, E. (1984). “A Simple Theory of Trade with Multinational Corporations.” Journal of Political Economy, Vol. 92, No. 3, pp. 451–47l. Kaufmann, D., A. Kraay, and M. Mastruzzi, (2003). Governance Matters III: Governance Indicators for 1996–2000. World Bank Policy Research Working Paper 4654. The World Bank. Kaufmann, D., A. Kraay, and M. Mastruzzi, (2008). Governance Matters VII: Governance Indicators for 1996–2007. World Bank Policy Research Working Paper 3106. The World Bank. King, G., and J. Honaker (2005). “What to do about Missing Values in Time Series Cross-Section Data.” Unpublished manuscript. King, Robert G. R. Levine (1993). “Finance, Entrepreneurship, and Growth: Theory and Evidence.” Journal of Monetary Economics, Vol. 32, No. 3, pp. 513–542. King, G., J. Honaker, A. Joseph, and K. Scheve (2001). “Analyzing Incomplete Political Science Data: An Alternative Algorithm for Multiple Imputation. American Political Science Review, Vol. 95, pp. 49–69. Kugler, T. and S. Swaminathan (2006). “The Politics of Population.” International Studies Review, Vol. 8, pp. 581–596. Jun, J. (1994). “How Taxation Affects Foreign Direct Investment.” International Economics Department, Debt and International Finance Division, The World Bank. La Porta, R., F. Lopez-de-Silanes, A., Shleifer, and R. Vishny (1997). “Legal Determinants of External Finance.” Journal of Finance, Vol. 52, No. 3, pp. 1131–1150. La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R. Vishny, (1998). “Law and Finance.” Journal of Political Economy, Vol. 106, No. 6, pp. 1113–1155. Lundstrom, S. (2005). “The Effect of Democracy on Different Categories of Economic Freedom. European Journal of Political Economy, Vol. 21, pp. 967–980. Mansfield, Edwin., and Romeo, A. (1980). “Technology Transfer to Overseas Subsidiary by US-based Firms.” Quarterly Journal of Economics, Vol. 95, No. 4, pp. 737–750. Markusen, J.R. (2001). Multinational Firms and the Theory of International Trade, Cambridge, MA: MIT Press. Maxfield, S. (1990). Governing Capital: International Finance and Mexican Politics, Ithaca: Cornell University Press. Mlambo, Kupukile (2006). “Reviving Foreign Direct Investments in Southern Africa: Constraints and Policies.” African Development Review, Vol. 17, Issue 3, pp. 552–557. Moran, Theordore H., E. M. Graham, and M. Blomström (2005). Does Foreign Direct Investment Promote Development? Washington, DC: Peterson Institute. Nigh, D. (1985). “The Effect of Political Events on United States Direct Foreign Investment: A Pooled Time-series Cross-sectional Analysis.” Journal of International Business Studies, Vol. 16, pp. 1–17.

Downloaded By: [Claremont Colleges] At: 01:37 14 December 2008

422

T. G. Coan and T. Kugler

Oman, Charles (2000). Development Centre Studies Policy Competition for Foreign Direct Investment. Paris: Organization for Economic Co-Operation & Development (OECD). Organski, A. (1965). The Stages of Political Development, New York: Alfred A. Knopf. Organski, A., J. Kugler, T. Johnson, and Y. Cohen, (1984). Births, Deaths, and Taxes: The Demographic and Political Transitions. Chicago: University of Chicago Press. Root, H. L. (1996). Small Countries, Big Lessons: Governance, and the Rise of East Asia, New York: Oxford University Press. Rubin, D.B. (1987). Multiple Imputation for Nonresponse in Surveys, New York: Wiley. Schmidt, G.W. (2003). Dynamics of Endogenous Growth Theory: A Case Study of the “Romer Model,” San Francisco: Elsevier Press. Schneider, F. and B. Frey (1985). “Economic and Political Determinants of Foreign Direct Investment.” World Development, Vol. 13, pp. 161–175. Steinberg, David A. and Stephen M. Saideman, (2008). “Laissez Fear: Assessing the Impact of Government Involvement in the Economy on Ethnic Violence.” International Studies Quarterly, Vol. 52, Issue 2, pp. 235–259. Tsai, P. (1994). “Determinants of Foreign Direct Investment and Its Impact on Economic Growth.” Journal of Economic Development, Vol. 19, pp. 137–163. Tumen, J.P. and C.F. Emmert (2004). “The Political Economy of U.S. Foreign Direct Investment in Latin America: A Reappraisal.” Latin American Research Review, Vol. 39, No. 3, pp. 9–28. United Nations Conference on Trade and Development (UNCTAD) (2007). World Investment Report. New York, NY. Wheeler, D. and A. Mody (1992). International Investment Location Decisions. Journal of International Economics, Vol. 33, pp. 57–76. Wei, S. (2000). How Taxing Is Corruption on International Investors? The Review of Economics and Statistics, Vol. 82, No. 1, pp. 1–11. The World Bank (2004). Doing Business in 2004. The World Bank Publications. The World Bank (2005). World Development Report 2005: A Better Investment Climate for Everyone. The World Bank Publications. The World Bank (2008). Doing Business in 2008. The World Bank Publications.